Arduino venting

[ralphd]

Although I usually code with vim and avr-gcc from the shell, I decided to modify the ArduinoISP code instead of writing an avrisp compatible programmer from scratch.  I found out that despite the updated SPI code in 1.6, there's still some dumb stuff going on.
https://github.com/arduino/Arduino/issues/3666

Rhetorical question: does Marchese Banzi know what a code review is?

[ralphd]

Theory: Banzi is a lousy programmer, and his delusions of grandeur keep him from realizing it.  Therefore some of the people he hires are just as clueless (and by chance a few are clueful).  They all have the ability to modify the Arduino core, and the one or two clueful people there are too "polite" to complain about working with morons...

[legacy]

no doubt he is a crap, he is also a complete ignorant about "debugging", when he launched Arduino/Zero, he was really thinking that (and he said that) - "to debug something you just need the serial print" - oh my goodness 

also, he appeared on a few RS videos arguing - "with Arduino, you can build interfaces for spaceships" - Really, why so serious, buddy ? 

[george graves]

So who wants to write a better IDE/core?  Any takers?

[shebu18]

Quote from: george graves on Today at 21:29:37

So who wants to write a better IDE/core?  Any takers?



>

Why that? There are enough alternatives to the arduino IDE

[tszaboo]

I stopped looking at Arduino libraries, when I realized they contain:

#include whatever.c


  <- this smiley 20 times.

[ralphd]

So who wants to write a better IDE/core?  Any takers?

Already been done, but Banzi's peeps are too clueless to look for something better.
www.wiring.org.co/
uecide.org/
https://github.com/mikaelpatel/Cosa

You can also use the atmel avr toolchain (i.e. avr-gcc & avr-libc) with Eclipse.

[ralphd]

no doubt he is a crap, he is also a complete ignorant about "debugging", when he launched Arduino/Zero, he was really thinking that (and he said that) - "to debug something you just need the serial print" - oh my goodness 

also, he appeared on a few RS videos arguing - "with Arduino, you can build interfaces for spaceships" - Really, why so serious, buddy ? 

And unfortunately most humans naively associate authority and fame with intelligence and ability.  i.e. He's the head of Arduino, so he's an expert on embedded computing.
But anyone who has seen how many people get their medical advice from Jenny McCarthy won't be surprised by Banzi's pack of lemmings.
http://dangerousminds.net/comments/even_c3po_and_r2d2_think_jenny_mccarthy_is_an_idiot

[ralphd]

I stopped looking at Arduino libraries, when I realized they contain:

#include whatever.c


  <- this smiley 20 times.

While I suspect when you see that in the Arduino libraries someone didn't realize what they are doing.
However I, reluctantly, have done that as a way to have pin definitions evaluated at compile time.
i.e. in main.c:

Code: [Select]

#define LCD_EN 2
#define LCD_RS 3
...
#define LCD_PORT PORTB

#include "lcd.c"

void main(void)
{
   lcd_print('x');
}

My preferred solution is to define configuration information like LCD_PORT in a Makefile, but when I did the #include "lcd.c" hack I didn't have a good build framework.  Now I use DMBS.
https://github.com/nerdralph/dmbs

[mikerj]

Although I usually code with vim and avr-gcc from the shell, I decided to modify the ArduinoISP code instead of writing an avrisp compatible programmer from scratch.  I found out that despite the updated SPI code in 1.6, there's still some dumb stuff going on.
https://github.com/arduino/Arduino/issues/3666

Rhetorical question: does Marchese Banzi know what a code review is?

Irrespective of the quality of the Arduino code (and I agree it's not great), your attitude stinks.  If you want the Arduino team to fix something, why do you think abusing them is either acceptable or likely to help?  It's not a professional or mature way to behave.

[ralphd]

Although I usually code with vim and avr-gcc from the shell, I decided to modify the ArduinoISP code instead of writing an avrisp compatible programmer from scratch.  I found out that despite the updated SPI code in 1.6, there's still some dumb stuff going on.
https://github.com/arduino/Arduino/issues/3666

Rhetorical question: does Marchese Banzi know what a code review is?

Irrespective of the quality of the Arduino code (and I agree it's not great), your attitude stinks.  If you want the Arduino team to fix something, why do you think abusing them is either acceptable or likely to help?  It's not a professional or mature way to behave.

I don't care if the Arduino core stays crappy.  I pointed out a couple bugs and got a dumb and condescending response, so screw them.
The eevblog forums are not for prudes and sissies.  You'd fit in better on the Arduino forums, or even avrfreaks.

[westfw]

Quote

there's still some dumb stuff going on.
>https://github.com/arduino/Arduino/issues/3666

"Pins left as output mode after SPI.end()" - that's a pretty mild "bug" to elicit such a vehement reaction.  I'd even call it a "design decision."

[ralphd]

Quote

there's still some dumb stuff going on.
>https://github.com/arduino/Arduino/issues/3666

"Pins left as output mode after SPI.end()" - that's a pretty mild "bug" to elicit such a vehement reaction.  I'd even call it a "design decision."

The bigger bug is the SS handling.  You are a smarter coder than the Arduino folks, so I'm sure if you read the begin() code you'll see how stupid it is.
And there's the library docs that are wrong.

At least they have enough clue to admit thepins being left as output is a bug instead of pulling the 'design intent' excuse...

[legacy]

One of the problems with Banzi is that he thinks the open source is always better, but,  trying to figure out his vision of "what is open", we discover he means that he is always open with other people's source code, especially sad with the customer care (that he gets almost for free)

[Bruce Abbott]

So who wants to write a better IDE/core?  Any takers?

Don't be silly. The only 'takers' here are those who are too lazy to write their own code - then complain when free stuff doesn't do what they think it should.



 

[ralphd]

One of the problems with Banzi is that he thinks the open source is always better, but,  trying to figure out his vision of "what is open", we discover he means that he is always open with other people's source code, especially sad with the customer care (that he gets almost for free)

And most people are ignorant of the fact that the Arduino core is a copy of Wiring, which is based on MITs Processing language.
Despite Banzi's pleas to support Ardino, he hasn't shared any profits with Hernando's team or MIT.
If I were going to support anyone it would be Sparkfun, who designed and open sourced basic boards like the pro mini.  Living outside of the US, the shipping cost from Sparkfun is prohibitive.

[ralphd]

So who wants to write a better IDE/core?  Any takers?

Don't be silly. The only 'takers' here are those who are too lazy to write their own code - then complain when free stuff doesn't do what they think it should.

Yup, I'm too lazy to write my own code.  Unless you count two bootloaders written in AVR assembler, a HD44780 library written in C, a bit-bang UART written in AVR assembler, C code (and a simple circuit) for doing time-division duplexed SPI with 2 pins, the fastest available bit-bang SPI master code for the AVR (both in C and assembler), ...

[Maxlor]

I don't get this arduino bashing. It's a tool that works a lot better than one'd think for the price. But if you're surprised when it fails to fulfill your specs as soon as they get detailled, or if you think it saves you from doing your job, you're delusional. Take it for what it is, something to build quick and dirty hacks with, and it's good for that. Once you get to the point where you're reviewing their libs for bugs (and... no, it's not a bug if you "feel" it should work differently) you're probably better off doing your project in a proper, controlled fashion.

[free_electron]

is';t the name a dead giveaway ? it's pronounced hardwhino ...

[jnz]

I don't get this arduino bashing. It's a tool that works a lot better than one'd think for the price. But if you're surprised when it fails to fulfill your specs as soon as they get detailled, or if you think it saves you from doing your job, you're delusional. Take it for what it is, something to build quick and dirty hacks with, and it's good for that. Once you get to the point where you're reviewing their libs for bugs (and... no, it's not a bug if you "feel" it should work differently) you're probably better off doing your project in a proper, controlled fashion.

Seemingly true.

If you need a micro for a professional project and are using Arduino, well, that's pretty much a telltale to me that something is wrong. I never thought people were basing real designs on them until I saw one show up in a product I was looking over.  Honestly, I'm just still pissed at myself for being on PIC for so many years before moving to ARM. I see anyone using Arduino now in a production environment as making that same mistake I was (granted there is some merit with the DSP if you really utilize it, but their tools and pricing are just so terrible I'd take a slower more expensive ARM in a heartbeat). Looking over Atmel's non-ARM line, I just don't get it.

[jancumps]

Although I usually code with vim and avr-gcc from the shell, I decided to modify the ArduinoISP code instead of writing an avrisp compatible programmer from scratch.  I found out that despite the updated SPI code in 1.6, there's still some dumb stuff going on.
https://github.com/arduino/Arduino/issues/3666

Rhetorical question: does Marchese Banzi know what a code review is?

what a poor issue you've logged there, man. Low.

[c4757p]

@facchinim You're only half right. SS does NOT have to be configured as output. One reasonable option is to enable the pullup, which the first part of SPI.begin() does... before setting it to OUTPUT.

From looking at your github account it seems you actually work for Arduino. I guess Marchese Banzi hires people with no more clue about writing good software than he does, hence the arduino core being the piece of shit it is...
... and I'm the fool for thinking I could help make it better...

So, is marching right into a bug report insisting the person responsible is an idiot and insulting the project typical behavior for you?

Does it work?

[zapta]

Although I usually code with vim and avr-gcc from the shell, I decided to modify the ArduinoISP code instead of writing an avrisp compatible programmer from scratch.  I found out that despite the updated SPI code in 1.6, there's still some dumb stuff going on.
https://github.com/arduino/Arduino/issues/3666

Rhetorical question: does Marchese Banzi know what a code review is?

what a poor issue you've logged there, man. Low.

Yes, not nice.

"From looking at your github account it seems you actually work for Arduino. I guess Marchese Banzi hires people with no more clue about writing good software than he does, hence the arduino core being the piece of shit it is...
... and I'm the fool for thinking I could help make it better..."

[ralphd]

@facchinim You're only half right. SS does NOT have to be configured as output. One reasonable option is to enable the pullup, which the first part of SPI.begin() does... before setting it to OUTPUT.

From looking at your github account it seems you actually work for Arduino. I guess Marchese Banzi hires people with no more clue about writing good software than he does, hence the arduino core being the piece of shit it is...
... and I'm the fool for thinking I could help make it better...

So, is marching right into a bug report insisting the person responsible is an idiot and insulting the project typical behavior for you?

Does it work?

What makes you think I want the Arduino project to fix their bugs?
Does Dave care if the Batterizer fixes their product?  No.

I don't have to kiss anyone's ass.

[ralphd]

Although I usually code with vim and avr-gcc from the shell, I decided to modify the ArduinoISP code instead of writing an avrisp compatible programmer from scratch.  I found out that despite the updated SPI code in 1.6, there's still some dumb stuff going on.
https://github.com/arduino/Arduino/issues/3666

Rhetorical question: does Marchese Banzi know what a code review is?

what a poor issue you've logged there, man. Low.

Were you too lazy to read the begin() code, or are you to dumb to see the bug?

[ralphd]

Although I usually code with vim and avr-gcc from the shell, I decided to modify the ArduinoISP code instead of writing an avrisp compatible programmer from scratch.  I found out that despite the updated SPI code in 1.6, there's still some dumb stuff going on.
https://github.com/arduino/Arduino/issues/3666

Rhetorical question: does Marchese Banzi know what a code review is?

what a poor issue you've logged there, man. Low.

Yes, not nice.

"From looking at your github account it seems you actually work for Arduino. I guess Marchese Banzi hires people with no more clue about writing good software than he does, hence the arduino core being the piece of shit it is...
... and I'm the fool for thinking I could help make it better..."

I'm not being nice?  Oh, no, nobody will like me and be my friend! Whatever will I do?

Sarcasm aside, when I want to be nice I do things like donate to the Against Malaria Foundation.
www.againstmalaria.com

[Nerull]

This thread is a brillant example of dunning-kruger.

Someone responds with a reasonable comment and you jump right to insults. Seems to me you're the sort of self-important jackass who drags most software projects down.

[ralphd]

This thread is a brillant example of dunning-kruger.

Someone responds with a reasonable comment and you jump right to insults. Seems to me you're the sort of self-important jackass who drags most software projects down.

I'm the unskilled one?  You talk big but where's your code?
Mine is at github.com/nerdralph
You'll also find several writeups on Hackaday.

You accuse me of resorting to insults while calling me a jackass.  Heard of hypocrisy?

[legacy]

One of the problems with Banzi is that he thinks the open source is always better, but,  trying to figure out his vision of "what is open", we discover he means that he is always open with other people's source code, especially sad with the customer care (that he gets almost for free)

And most people are ignorant of the fact that the Arduino core is a copy of Wiring, which is based on MITs Processing language.
Despite Banzi's pleas to support Ardino, he hasn't shared any profits with Hernando's team or MIT.
If I were going to support anyone it would be Sparkfun, who designed and open sourced basic boards like the pro mini.  Living outside of the US, the shipping cost from Sparkfun is prohibitive.

Exactly, but I want to add AdaFruit Industries to the list because I am happy to support them because they have done a lot of useful things.
Helpful collaboration.

Living outside of the US, the shipping cost from Sparkfun is prohibitive.

unfortunately, I live in Italy, It's extremely expensive for me, but I have to thanks Sparkfun for their contribution with the Eagle CAD libraries. A lot of useful things come from them. AdaFruit Industries has done a few things I like, including their Eagle CAD libraries and a few good LCD kits for cheap.

[c4757p]

This thread is a brillant example of dunning-kruger.

Someone responds with a reasonable comment and you jump right to insults. Seems to me you're the sort of self-important jackass who drags most software projects down.

I'm the unskilled one?  You talk big but where's your code?

Actually, it sounds more like English might not be your native language. Is that true? He never called you unskilled.

[ralphd]

This thread is a brillant example of dunning-kruger.

Someone responds with a reasonable comment and you jump right to insults. Seems to me you're the sort of self-important jackass who drags most software projects down.

I'm the unskilled one?  You talk big but where's your code?

Actually, it sounds more like English might not be your native language. Is that true? He never called you unskilled.

It sounds like you don't know what dunning-krudger means.  Is that true.
FWIW, English is my mother tounge, mais je parle français aussi.

[c4757p]

Fair enough, though I'd have assumed the "dunning-krudger" reference here to possibly indicate that you falsely believe you're in a position to respond the way you did.

Either way, you seem like you are likely supremely unpleasant to work with.

[ralphd]

Fair enough, though I'd have assumed the "dunning-krudger" reference here to possibly indicate that you falsely believe you're in a position to respond the way you did.

Either way, you seem like you are likely supremely unpleasant to work with.

Yes, incompetent fools would likely not enjoy working with me.  Hence I rarely have to put up with such people.
If you are unskilled and eager to learn, I'll gladly share my knowledge.  I've taught hundreds of students Java, C++, Unix, and C.
And I respond to all non-spam questions on my blog http://nerdralph.blogspot.com.

[westfw]

'm sure if you read the begin() code you'll see how stupid it is.

I didn't look at the begin() code.   the bug report didn't explicitly mention any bug in begin().
Like the responses in the bug report said, the AVR has "weird" handling of SS; setting it as output doesn't seem too awful to me...

So who wants to write a better IDE/core?  Any takers?

Sure.
Submit it to the Arduino folk to include in their system for free?  OK.
Maintain it for the next 10 or 20 years, responding to anyone who thinks that they can improve some odd edge case?  Putting up with insults chasing compiler and hardware changes?   No way.  Optiboot is bad enough. :-)

And most people are ignorant of the fact that the Arduino core is a copy of Wiring [by Hernando Barragán], which is based on MITs Processing language.

Wiring is a fine example of a rather abandoned project :-(  In the last several years, it's gotten only a smatter of minor bugfixes, despite early promises of "cross platform support for xmega, msp430, pic32, and ARM cores."  And none of them by Hernando (actually, it doesn't look like Hernando was particularly involved in the initial publication, either.)
   If you want to believe that it's "better than arduino" because it got some improvements that the arduino team refused, 4 years ago, that's fine.  (You may even be right, since some of the things that Arduino has added have bloated the older platforms.)  If you ask me, Hernando never had any interest in commercializing either Arduino or Wiring (the man is an Artist! (based on his web page)) (which is fine.  I've never heard him complain about Arduino, either.)
I don't think that "wiring" is based on "processing" except for the IDE.  (Have you ever used Processing?)  (and Processing isn't a "language" any more than "Arduino" is; it's Java with preprocessing and libraries...)

[ez24]

From an outsider, it is hard to tell who is insulting who.  How about use names.  PIC is better right?

[ralphd]

'm sure if you read the begin() code you'll see how stupid it is.

I didn't look at the begin() code.   the bug report didn't explicitly mention any bug in begin().

I think age is catching up on you Bill.
"If SS is in input mode before calling SPI.begin(), it is left as output high."

Like the responses in the bug report said, the AVR has "weird" handling of SS; setting it as output doesn't seem too awful to me...

But setting it to output when the docs imply that control of SS is left to the user code is bad.  Sometimes setting it to output low and other times setting it to output high is pretty awful to me.  But as I said, if you look at the begin() code, you'll see what I'm talking about.

I am actually a bit surprised by your comments.  I know you've looked at my code before, and therefore you know I am skilled in AVR programming.  For AVR assembler, I have yet to encounter anyone better than me, and I only bought my first AVR 2 years ago.  I recall you saying you tried to re-write optiboot in assembler to get it <= 256 bytes (minimum bootloader size on the mega8 & mega168), but couldn't do it.  PicobootSTK500 implements the same susbset of STK500 that optiboot does, but in 224 bytes, and it includes eeprom reading.  One of these days when I have a free afternoon I'll add eeprom write support, which will bump it up to ~240 bytes.

p.s. You previously mentioned that Arduino doesn't support reading/writing the eeprom.  It looks like it does, at least from 1.6; the build log shows it creates a .eep file for eeprom data.

[ralphd]

I don't think that "wiring" is based on "processing" except for the IDE.  (Have you ever used Processing?)  (and Processing isn't a "language" any more than "Arduino" is; it's Java with preprocessing and libraries...)

It's not a personal observation, it's what the Wiring folks say:
"Wiring builds on Processing, an open project initiated by Casey Reas and Ben Fry, both formerly of the Aesthetics and Computation Group at the MIT Media Lab."
http://wiring.org.co/about.html

I agree "language" may not be the best word; I did pause for a moment when I wrote it but a better word didn't come to me.  Perhaps framework may be a better term.

[legacy]

The Dunning–Kruger  effect is a cognitive bias wherein relatively unskilled individuals suffer from illusory superiority, mistakenly assessing their ability to be much higher than is accurate. This bias is attributed to a metacognitive inability of the unskilled to accurately evaluate their own ability level. Conversely, highly skilled individuals may underestimate their relative competence, erroneously assuming that tasks that are easy for them are also easy for others.

[…]

I wanted to know the exact meaning of a new word (English is my second language, still learning) and it banged a laugh.

I have read that "Dunning–Kruger" is a very bloody expression, context dependent, sometimes sarcastic, so not really more offensive than how bloody it sounds, but ... to be used carefully because it may flame bonfires.

[IconicPCB]

This

"...   For AVR assembler, I have yet to encounter anyone better than me, and I only bought my first AVR 2 years ago. ..."

 reminds me of FIGJAM.

[amyk]

About the other pins not being restored to the original state after end() is called, we'd better try to understand what people want to do with that pins after calling end(), so maybe it's better to open a thread on the mailing list first.

I agree this guy is a bit dense, putting the pins back the way they were seems the most logical choice. It doesn't require understanding anything...

[ez24]

Dunning–Kruger

FIGJAM

English is my language (at least because I was born in the USA) and I am learning a lot of words on this forum - I did not know those words.  And every time I see the title of this post I think this is about Arduino exploding because of venting problems like caps.  Still getting use to the boxing here. 

[westfw]

you know I am skilled in AVR programming.

Yep.  Unfortunately, that doesn't have much to do with being able to design a "system" for the Arduino target audience.
Some of the design decisions in Arduino are brilliant.  Some are ... frustrating.   Some are both.  (for instance, considering the number of supposedly "somewhat experienced" programmers I've seen who can't quite seem to understand bitwise operations, the whole "pin number" thing was a great idea.  It's also frustratingly slow and simple-minded.)  If you say "setting SS as output should be left to the user" and the Arduino team says "Too many users would get it wrong", I'm going to have to side with the Arduino team.  A fair percentage of arduino users don't remember to set the pins to outputs when they are explicitly using them as outputs ("Why is my LED so dim?")

You previously mentioned that Arduino doesn't support reading/writing the eeprom.  It looks like it does, at least from 1.6;

I think it has always generated the .eep files, but it doesn't pass them to avrdude, and of course optiboot doesn't support receiving them.


BTW, I notice that Wiring has the same SPI "bug" - how is it going, getting them to change it?

There's an ASF SPI Master example for ATmega128RFA1; it doesn't seem to do anything with SS in either the ASF SPI library (that I can tell) or in the "user" code.  I guess it's targeted at a particular eval board, but ... I wonder if it works right?

[zapta]

...Sarcasm aside, when I want to be nice I do things like donate to the Against Malaria Foundation.
www.againstmalaria.com

"... but sometimes I am not a jerk" is a funny excuse.

Massimo Banzi and the rest of the Arduino team did the impossible, bringing C/C++ and embedded programming to the masses and on the way creating a huge eco system of open source and affordable controllers and peripherals. For that reason they have my great respect even if they chose to design some aspects differently than I would.

[ralphd]

...Sarcasm aside, when I want to be nice I do things like donate to the Against Malaria Foundation.
www.againstmalaria.com

"... but sometimes I am not a jerk" is a funny excuse.

As I said before I don't give a fuck about anyone's subjective emotional evaluation of my personality.  I do care about making a positive impact on the world by objective standards.  How many lives do you save by being polite when people say dumb things while claiming to be smart?
http://www.thelifeyoucansave.org/

Massimo Banzi and the rest of the Arduino team did the impossible, bringing C/C++ and embedded programming to the masses and on the way creating a huge eco system of open source and affordable controllers and peripherals. For that reason they have my great respect even if they chose to design some aspects differently than I would.

And what evidence do you have that Arduino was so successful because of Banzi and not in spite of him?

I respect people with talent, not luck.  So I respect people like Adilson Akashi (who you probably never even heard of since you're so focused on Banzi's BS).  I even have some respect for David Mellis; the code he wrote for Arduino was OK for a grad student, but he never claimed he was the next Bjarne Stroustrup or James Gosling.  However I distaste people who take credit for other people's work the way Banzi does.

[legacy]

Banzi has stolen the idea from a thesis of one of his student, this is how the story has begun, then he has recycled ideas from MIT, effort from other people, and, in business with Atmel, he continuously claims that HE has invented the wheel 

indeed, probably Banzi can also master the Neanderthal Technology, claiming HE has invented everything in the human technology - I discovered fire, invented wheel, built arduino, created a marvelous and high productive IDE, and introduced C++ to embedded, anything was existing before me, and I say you have the power on your finger tips just with a serial.print, let me show you a (wrong ever) tutorial to learn how - in the last 28000 years (because 28000 years ago he hit with his bat a Neanderthal Guru in order to steal all the rocks and stones with signs on them)

Go Banzi! 

[zapta]

Go Banzi! 

Which brings to mind the best TV show ever

[Bruce Abbott]

(and he said that) - "to debug something you just need the serial print" - oh my goodness 

'serial print'? 'debug'? If my code doesn't work I just go back to the source and fix it. 

Since the Arduino is already connected to the PC via a serial port, wouldn't 'serial print' be the obvious way to debug it?

Arduino was designed for people who just want to write a routine and get it running with minimum hassle. No messing around with compilers and linkers, programmers, in circuit debuggers etc. It even creates the main loop for you. This is not a platform for 'serious' developers...


 


 
 
 

[ralphd]

Banzi has stolen the idea from a thesis of one of his student,

I think most people who have gone to a university that has post-graduate studies is familiar with professors that succeed by finding good students and taking credit for their ideas.

Looking at Hernando's thesis you can see that Banzi was just his advisor.  As is typical in a thesis, the student will try to reference the advisor's previous work, and hence the short mention of programma2003 (which is just a PCB for previously published PIC circuits: http://www.voti.nl/blink/index.html).
http://wiki.wiring.co/images/7/76/Wiring_thesis.pdf

Back in 2005, before the success of the Arduino project had gone to his head, Banzi seemed to be giving proper credit to people like Dave Mellis:
https://web.archive.org/web/20050305232506/http://www.potemkin.org/

And it's pretty clear that Dave Cuartielles deserves most of the credit for evolving Hernando's Wiring board into the Arduino (with Dave Mellis contributing as well).
http://mazine.ws/node/195

Even by 2006 Dave C. was still getting credit for the Arduino board with no mention of Banzi:
https://web.archive.org/web/20060506014621/http://www.0j0.org/wiki/index.php/DojoDave

From 2006 to 2009 Dave M, Dave C, and Adilson Akashi clearly made contributions to the Arduino hardware and software, but I could find no record of Banzi making any technical contributions.
https://www.arduino.cc/en/uploads/Main/ArduinoSeverinoManual2.pdf
http://code.google.com/p/arduino/source/browse/tags/0015/hardware/bootloaders/atmega/ATmegaBOOT_168.c

p.s. Interesting tidbit about Dave C: he got into computer hardware/software with a C64 as a kid.

[retrolefty]

(and he said that) - "to debug something you just need the serial print" - oh my goodness 

'serial print'? 'debug'? If my code doesn't work I just go back to the source and fix it. 

Since the Arduino is already connected to the PC via a serial port, wouldn't 'serial print' be the obvious way to debug it?

Arduino was designed for people who just want to write a routine and get it running with minimum hassle. No messing around with compilers and linkers, programmers, in circuit debuggers etc. It even creates the main loop for you. This is not a platform for 'serious' developers...

I agree, now micro based projects can include 'serious' hobbyists, 'serious' artists, and'serious' Ham Radio operators. 

 Opening microcontroller development for the masses seems to often bring scorn from some elitists. Now not all have to pay the same 'dues' to gain fun, fame, and happiness in their specific endeavors.  Gate-keeping is so 20th century.
The Arduino developers are neither Saints nor Gods, just a group that happen to have a good idea at a good time for a new user base.

[westfw]

take credit for other people's work the way Banzi does.

This seems to be at the core of your distaste for Arduino.   But ... exactly where does this happen?  I mean, Banzi is definitely the "spokes-model" and main evangelist for Arduino, but I don't recall ever hearing him take much credit for development.  His videos contain a lot of "we", and his presence on the Arduino site is relatively understated.  "Founder" and "co-developer of the original board."

CEOs and Founders frequently get more credit than they deserve, once a startup becomes a significant-sized business.  And they take more of the blame.
That's part of their job.   Like most engineers, I'm pretty introverted, and ... I wouldn't want that job for all the glory in the world.
Everyone has heard (?) of John Chambers, but few people recognize names like Bill Graves, Chuck Sutcliff, David Means, or Kirk Lougheed (or Bill Yeager, since we want to include academics whose seminal ideas were (maybe, sort-of) stolen.)

[Bruce Abbott]

it happened when Banzi announced the new Arduino/Zero (arm), which has a debug chip installed,

Arduino Zero

The Atmel Embedded Debugger (EDBG) implements a JTAG interface in order to program the on-board SAMD21 and is also connected to hardware serial of the microcontroller. This means that the 'Serial' class responds to the programming port of the board. The Arduino Zero has been designed in collaboration with ATMEL, and the on-board EDBG can be used through ATMEL Studio to get full access to the microcontroller memories to help debug your code.

So you 'serious' developers can debug the Arduino Zero to your heart's content, while the rest of us get on with just doing stuff.

Quite the joke is it really. My PC has a 2.8GHz Pentium D processor, but apparently this is far too slow because it takes over 2 minutes to load Atmel studio 6 - by which time I am bored and decide to do something else instead. Compare that to the Arduino IDE which loads in 15 seconds. I could have written a program, downloaded it into the Arduino and had it running before Atmel's bloatware even gets started. 

But what do I know? I'm just an old curmudgeon who's been coding micros since 1980, not a serious programmer...   

[Bassman59]

Quite the joke is it really. My PC has a 2.8GHz Pentium D processor, but apparently this is far too slow because it takes over 2 minutes to load Atmel studio 6 - by which time I am bored and decide to do something else instead.

Preach it, brother! I've got a 3-ish GHz quad Core i7 machine with 16 GB RAM running 64-bit Windows 7 and Atmel Studio 6 (still 32 bit? Fuck that) takes forever to load. I spend the time waiting for it to load looking for alternatives. So far the GNU ARM Eclipse plug-ins looks very promising. (If only the SAM-ICE I bought wasn't Yet Another Proprietary Fucking JTAG Dongle.)

But what do I know? I'm just an old curmudgeon who's been coding micros since 1980, not a serious programmer... 

Kids today, they don't know the pain (and the expense) of the Nohau bond-out pod 8051 emulator, with AMONG OTHER ANNOYANCES was its inability to do MOVX with an 8-bit address (something that I've never seen a compiler support, but handy from assembly if you use P2 for port I/O). Bonus: the pod always blew up the afternoon before the product was supposed to go out the door.

[Muxr]

(and he said that) - "to debug something you just need the serial print" - oh my goodness 

'serial print'? 'debug'? If my code doesn't work I just go back to the source and fix it. 

Since the Arduino is already connected to the PC via a serial port, wouldn't 'serial print' be the obvious way to debug it?

Arduino was designed for people who just want to write a routine and get it running with minimum hassle. No messing around with compilers and linkers, programmers, in circuit debuggers etc. It even creates the main loop for you. This is not a platform for 'serious' developers...

I agree, now micro based projects can include 'serious' hobbyists, 'serious' artists, and'serious' Ham Radio operators. 

 Opening microcontroller development for the masses seems to often bring scorn from some elitists. Now not all have to pay the same 'dues' to gain fun, fame, and happiness in their specific endeavors.  Gate-keeping is so 20th century.
The Arduino developers are neither Saints nor Gods, just a group that happen to have a good idea at a good time for a new user base.

+1

Yup. It's pretty easy to criticize other people's work. Heck i've been programming for the past 30 years and I am pretty sure you can find some jank code in most of my projects that could use refactoring.

The thing is my best code never gets shipped, but the code that gets written in a hurry under crazy deadlines always does. Such is the nature of what we do I suppose.

I certainly appreciate what Arduino folks accomplished, because for all its technical flaws Arduino is still the best beginner uController environment around. Their focus was clearly on the accessibility and not on the highest industrial robustness of their implementation.

[gmb42]

Quite the joke is it really. My PC has a 2.8GHz Pentium D processor, but apparently this is far too slow because it takes over 2 minutes to load Atmel studio 6 - by which time I am bored and decide to do something else instead.

But what do I know? I'm just an old curmudgeon who's been coding micros since 1980, not a serious programmer... 

Hmm, a 2.8GHz Pentium D was the slowest Pentium available when initially manufactured in 2005 or 2006, I would think that you're not Atmel's target market.  For comparison, using a 2 yr old i7-4770K with an SSD, Atmel 6.2 opens a (not insignificant) project in under 10 seconds.

No one is forcing you to upgrade, but a more modern system would be much nicer to use in all sorts of ways.

[c4757p]

Running an IDE is normally one of the lighter tasks a computer can do..... there's not much more in Atmel Studio that has to be loaded immediately (options deep in menus that can be loaded when requested need not apply) that is larger than the VB6 I recall running before 2000!

Lots of people have to use IDEs on older computers. My capstone project team at school last year switched to a pure arm-none-eabi-gcc solution because Atmel Studio was unusable on the machine the school gave us. I wouldn't be surprised to find many places of work in similar positions. That computer was not all too old... it ran matlab quite fast as I recall... Atmel Studio is extremely bloated as IDEs go.

[mikerj]

The eevblog forums are not for prudes and sissies.

The eevblog is for amateurs and professionals who enjoy helping each other solve problems.  It's not here for people to behave like spoilt children.  Grow up and behave yourself.

[ralphd]

The eevblog forums are not for prudes and sissies.

The eevblog is for amateurs and professionals who enjoy helping each other solve problems.  It's not here for people to behave like spoilt children.  Grow up and behave yourself.

So that puts you in the condesending prude category to think you control how I behave.
Go whine to Dave and see how far that gets you.

[Bruce Abbott]

Hmm, a 2.8GHz Pentium D was the slowest Pentium available when initially manufactured in 2005 or 2006,

Yeah right.

Tom's Hardware: Mother of All CPU Charts 2005/2006

For comparison, using a 2 yr old i7-4770K with an SSD, Atmel 6.2 opens a (not insignificant) project in under 10 seconds.

Not everybody can afford (or wants) to upgrade their PC every 2 years. But anyway I was talking about the time it takes for the IDE itself to load, not how long it takes to open a project.

[ralphd]

take credit for other people's work the way Banzi does.

This seems to be at the core of your distaste for Arduino.

No, that's at the core of my distaste for Banzi.  My distaste for the Arduino core comes from reading the source.  There are times I have cringed from reading code that is so bad.
For comparison, avr-libc, although it hasn't been bug-free, is generally well written.  A few things in it are sub-optimal, and some things are done in a different way than I would but still fall within the scope of acceptable software design.

[ralphd]

Wiring is a fine example of a rather abandoned project :-(

Don't count it out yet.  I have first-hand information from someone claiming to be on the development team saying they are still working on Wiring, and Hernando is still involved.
And although I won't disclose personal communications, I will say that my perception is that the team is at least confused by Banzi, and probably annoyed.

[gmb42]

Hmm, a 2.8GHz Pentium D was the slowest Pentium available when initially manufactured in 2005 or 2006,

Yeah right.

Tom's Hardware: Mother of All CPU Charts 2005/2006

Note sure what this is referencing, I used the font of all knowledge, WikiPedia looking at Pentium D.  Further checking with Intel in their archive for "Legacy Pentium Processors" gives this for Presler and this for Smithfield, and you're right, they did make a slower one, the mutant offspring 805 with a glorius 2.66GHz clock.  My point was that a 2.8GHz CPU was (almost) the slowest Pentium D made.  Yes it was faster than older CPUs, but it is definitely showing it's age now.

For comparison, using a 2 yr old i7-4770K with an SSD, Atmel 6.2 opens a (not insignificant) project in under 10 seconds.

Not everybody can afford (or wants) to upgrade their PC every 2 years. But anyway I was talking about the time it takes for the IDE itself to load, not how long it takes to open a project.

My info wan't clear, that time includes starting the IDE.  I double clicked the solution file in Explorer.

Regardless, my point wasn't to boast, but to show that attempting to castigate a piece of software for poor performance on a CPU nearly a decade old is slightly unfair IMO.

[Bruce Abbott]

Hmm, a 2.8GHz Pentium D was the slowest available when initially manufactured in 2005 or 2006...

My point was that a 2.8GHz CPU was (almost) the slowest Pentium D made.

Moving the goalposts? The Pentium D was most certainly not the slowest CPU around in 2005 - it was actually one of the fastest. To then turn around and label the 2.8GHz version as 'almost the slowest' simply because the top speed grade was 3.2GHz is disingenuous.

Now consider that Atmel Studio 6 was released in February 2002.  What was the fastest CPU you could get back then? Some of us remember those times, and we are sick of being told to buy faster hardware just so software can become even more bloated.

my point wasn't to boast, but to show that attempting to castigate a piece of software for poor performance on a CPU nearly a decade old is slightly unfair

Your system has a top-of-the-line CPU and a solid state hard drive. I bet it cost plenty and is worth boasting about. So boast away, we don't mind. 

What I do mind is expecting hobbyists, students and others with limited funds to spend big bucks on a more powerful PC, just so they can enjoy debugging their Arduino at a reasonable speed. What's the point of releasing a cheap dev board that doesn't need any extra debugging hardware, if you then have to spend hundreds (or even thousands) of dollars to make your PC powerful enough to use it? The answer is:- you shouldn't have to. That's why debugging via the serial port is still a valid option.

[NiHaoMike]

If you need a micro for a professional project and are using Arduino, well, that's pretty much a telltale to me that something is wrong. I never thought people were basing real designs on them until I saw one show up in a product I was looking over.

Even though I mostly work with PICs, I did a design for work with an embedded Arduino Leonardo to make it easier for others to maintain the code as time goes on. I have also come across many other internally developed tools (e.g. relay boards and thermal diode readers) that were based on Arduino.

[westfw]

consider that Atmel Studio 6 was released in February 2002

Um.  No.  2012, perhaps?  http://www.atmel.com/tools/STUDIOARCHIVE.aspx

Fortunately, one tends to spend so much time inside an IDE that its startup time is pretty irrelevant. One DOES wonder what it is that AS is doing to be so slow on "typical" systems, though.

[matseng]

For comparison, using a 2 yr old i7-4770K with an SSD, Atmel 6.2 opens a (not insignificant) project in under 10 seconds.

Not everybody can afford (or wants) to upgrade their PC every 2 years. But anyway I was talking about the time it takes for the IDE itself to load, not how long it takes to open a project.

I currently do all my development for ARM, golang and nodejs on a 2013 MacBook. Atmel Studio boot up (under Win7 on a Parallels VM in OSX) in less than 13 seconds. It is kinda irritating, but as said here previously you only start it like once a day so it really doesn't matter.  But I must admit that I still look back in awe at the good old stuff like Turbo Pascal that fitted on a 360 KB floppy, was a complete IDE with debugger, started up in a few seconds and compiled the code like greased lightning even on a 4.77 MHz 8088.